- Home
- Research
- Research Overall Objectives
- Objective 5
Objective 5: Integrated modelling of coupled human-natural systems
To be relevant to policy and decision making, scientific tools must consider the linkages between people and nature – termed “coupled human-natural systems.” In today’s world, economic challenges cannot be addressed without considering environmental change, drivers, and feedback – nor can environmental challenges be isolated from economic ones. ARIES (ARtificial Intelligence for Environment & Sustainability, http://aries.integratedmodelling.org/), an application of semantically integrated modelling technology using Artificial Intelligence (AI), integrates scientific data and models that simulate and integrate environmental and socioeconomic systems, deepening our understanding of the natural world and of how the choices society makes can impact future economic prosperity and environmental sustainability. It is conceived as a powerful networked software technology that allows to map natural capital, natural processes, human beneficiaries, and service flows to society. it has the potential to build in all the agents involved in the nature-society interaction, connect them into a flow network, and create for each agent the best possible models that simulate and integrate environmental and socioeconomic systems, deepening our understanding of the natural world and of how the choices society makes can impact future economic prosperity and environmental sustainability. The diffusion of the underlaying technology offers six critically needed advantages to 21st century interdisciplinary science: ability to combine products and workflows that would be difficult for individual humans due to their complexity; integration of different modelling paradigms from simple (e.g., deterministic and probabilistic models) to complex approaches (e.g., agent-based and networks) depending on context and scale; ability to operate smartly across scales, from local to global; flexible incorporation of the best available knowledge, from curated global public datasets to “big data” to user-provided data; adoption of non-ambiguous common scientific languages in both the implementation and delivery of products; tracking of quality and uncertainty throughout modelling workflows. The main activities under SO5 are as follow:
Activity 5.1. Scaled complexity in biophysical and social modelling.
The technology underlying ARIES can choose and assemble the most appropriate data and model components according to scale, detail, and availability of data. For example, in data-poor contexts, the system adopts Bayesian belief network models that approximate the likely outcomes of physical processes through causal paths and correlations learned from data in comparable contexts. When data is available, the system adopts spatially explicit dynamic models of different processes. This scaled complexity approach allows best response and explicit communication of uncertainty in diverse contexts. The technology is networked and distributed so that data and models can be served from any node connected to the ARIES semantic web. While most of the global dataset will be served, for default analysis, from the Basque nodes and physical infrastructure, the network is set to grow in a decentralized fashion whereby local and regional data authorities (e.g. National Statistical Offices) can put online and manage their own nodes. Further development and scaling up of ARIES and its components are foreseen partially subjected to IKUR support (upgrade capacity to support users worldwide is scale up to 3,000 users by 2023 and by 10,000 users by 2026; and yearly training courses for national and international students).
Activity 5.2. Bridging disciplines: from biophysical to social through agriculture and food security.
The investigation on social agents is synergic with many integrated scenarios developed in other outcomes. In order to link data and models produced by other communities (such as agricultural, hydrological, pedological and geological) ARIES uses an approach driven by semantics that enables the annotation of data and models of diverse provenance (non-semantic resources) through a declarative language and logical expressions (i.e. ontologies) that can be understood by machines while reusing endorsed vocabularies and terminologies from recognized institutions. In such way resources become Findable, Accessible, Interoperable andReusable (FAIR) through purely logical queries, which resolve into integrated models through machine reasoning and logical inference. Negotiating effective translation of terminologies and continued interoperability is a crucial prerequisite to developing any effective trans-disciplinary project. The planned modelling applications include further understanding of soil-hydrology-vegetation -livestock interactions, for the management of agro-silvopastoral systems in collaboration with RL3.
Activity 5.3. Bridging scales: from process detail and agent behaviour to economic and policy instruments.
Focusing on both, the biophysical mechanisms of ecosystem service provision, the socioeconomic drivers of their change and their implications for human well-being, requires integrating scales from local, through regional decision-making, to the country-level accounting of natural capital, which ultimately impacts the global economic system. Case studies for ARIES often cut across wide scale spans in space and time. Such activities have prompted the development of scale-aware methods and automatic scaling of models, with explicit account of the resulting uncertainties. Techniques such as Multiple Criteria Analysis have also been integrated into ARIES to provide rapid and easily communicable analysis and visualization of stakeholder priorities and potential conflicts over alternative goals at different scales. These instruments will allow the investigation of scaling trade-offs, such as those between short-term increase in provisioning services, and the long-term loss of regulating ones (e.g. flood risk). One case study that will be tackled is the modelling of human migration as consequence of environmental change, using machine learning and agent-based modelling techniques (Led by Stefano Balbi).
Activity 5.4. Building and delivering applications online.
Reconciling the need for simplicity and intuitiveness with that for accuracy, specificity and dynamic resolution is a challenge. By adapting models to diverse social, economic, and policy contexts without overly complicating their application by decision makers, ARIES opens the highly specialized practice of modelling to non-specialists, reduces its costs and potentially expands its role in decision making. ARIES can be used by decision makers to target economic incentives, while minimizing the information costs of identifying trade-offs between efficient and equitable targeting of ES providers. Several applications are already fully operational and other potential applications are under development: ecosystems services modelling access (for regional/sub-regional applicability – i.e. Basque Country, Cantabria); natural capital accounting and ARIES for SEEA (System of Environmental Economic Accounting - UNSD/UNEP collaboration); biodiversity (Post2020 Biodiversity framework) and SDGs indicators; conservation and restoration of critical social-ecological systems; connection of global funding sources with potential high-impact projects located in the highest priority areas defined by conservation organizations, scientists and governments (pilot project initially addressing Colombian Amazonia Forest but with the plan to scale-up over time) - IDB, ESRI, Salesforce collaboration; relevant databases semantically annotated and connected to ARIES (e.g. GEOEUSKADI).